Unverified50% confidenceBenchmarkExact time
Compared to pure decoder architectures (such as GPT), GLM achieves an average improvement of 8–15 percentage points on Chinese NLU tasks; compared to encoder-decoder architectures (such as T5), inference efficiency is improved by approximately 40%.
1
Sources
50%
Confidence
Long-term
Relevance
9/5/2026
First Seen
Sources
Related Entities
Related Claims
Unverified许多推理服务的GPU利用率不足30%，通过VGPU技术可将利用率提升至70%以上66% similarUnverifiedZhipu released GLM-5.3, claiming its coding capability improved by 50% compared to GLM-5.2.66% similarUnverifiedSelf-Consistency can bring improvements of 5–20 percentage points on mathematical reasoning benchmarks such as GSM8K.64% similarUnverified权重绑定在中小规模模型上(如GPT-2 124M)通常能提升性能并减少约30%的嵌入相关参数，但在超大规模模型中收益递减64% similarUnverified在 GSM8K 等数学基准上，PAL 和 PoT 相比纯思维链可将准确率提升 10-20 个百分点64% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/861037API
curl https://kongchang.com/api/v1/knowledge/claims/861037MCP
get_claim(id=861037)